RESEARCH
open access
Trajectories of risk after hospitalization for heart failure, acute myocardial infarction, or pneumonia: retrospective cohort study Kumar Dharmarajan,1 Angela F Hsieh,2 Vivek T Kulkarni,2 Zhenqiu Lin,2 Joseph S Ross,3 Leora I Horwitz,3 Nancy Kim,3 Lisa G Suter,4 Haiqun Lin,5 Sharon-Lise T Normand,6 Harlan M Krumholz,2
1Department
of Internal Medicine, Columbia University Medical Center, NY, USA 2Center for Outcomes Research and Evaluation, Yale-New Haven Hospital, New Haven, CT, USA 3Section of General Internal Medicine, Yale University School of Medicine, New Haven, CT, USA 4Section of Rheumatology, Yale University School of Medicine, New Haven, CT, USA 5Department of Biostatistics, Yale University School of Public Health, New Haven, CT, USA 6Department of Biostatistics, Harvard School of Public Health, Boston, MA, USA Correspondence to: K Dharmarajan, Section of Cardiovascular Medicine, Yale University School of Medicine, New Haven, CT 06510, USA
[email protected] Additional material is published online only. To view please visit the journal online (http:// dx.doi.org/10.1136/BMJ.h411) Cite this as: BMJ 2015;350:h411 doi: 10.1136/bmj.h411
Accepted: 17 December 2014
Abstract Objective To characterize the absolute risks for older patients of readmission to hospital and death in the year after hospitalization for heart failure, acute myocardial infarction, or pneumonia. Design Retrospective cohort study. Setting 4767 hospitals caring for Medicare fee for service beneficiaries in the United States, 2008–10. Participants More than 3 million Medicare fee for service beneficiaries, aged 65 years or more, surviving hospitalization for heart failure, acute myocardial infarction, or pneumonia.
67.4% and 35.8% of hospitalizations for heart failure, 49.9% and 25.1% for acute myocardial infarction, and 55.6% and 31.1% for pneumonia. Risk of first readmission had declined 50% by day 38 after hospitalization for heart failure, day 13 after hospitalization for acute myocardial infarction, and day 25 after hospitalization for pneumonia; risk of death declined 50% by day 11, 6, and 10, respectively. Daily change in risk of first readmission to hospital declined 95% by day 45, 38, and 45; daily change in risk of death declined 95% by day 21, 19, and 21. After hospitalization for heart failure, acute myocardial infarction, or pneumonia, the magnitude of the relative risk for hospital admission over the first 90 days was 8, 6, and 6 times greater than that of the general older population; the relative risk of death was 11, 8, and 10 times greater.
Main outcome measures Daily absolute risks of first readmission to hospital and death for one year after discharge. To illustrate risk trajectories, we identified the time required for risks of readmission to hospital and death to decline 50% from maximum values after discharge; the time required for risks to approach plateau periods of minimal day to day change, defined as 95% reductions in daily changes in risk from maximum daily declines after discharge; and the extent to which risks are higher among patients recently discharged from hospital compared with the general elderly population.
Conclusions Risk declines slowly for older patients after hospitalization for heart failure, acute myocardial infarction, or pneumonia and is increased for months. Specific risk trajectories vary by discharge diagnosis and outcome. Patients should remain vigilant for deterioration in health for an extended time after discharge. Health providers can use knowledge of absolute risks and their changes over time to better align interventions designed to reduce adverse outcomes after discharge with the highest risk periods for patients.
Results Within one year of hospital discharge, readmission to hospital and death, respectively, occurred following
Introduction Patients are vulnerable to major adverse outcomes after hospital stay. Readmissions are common: nearly one in five adults aged more than 65 is readmitted to hospital within 30 days of discharge.1 Death is also common in this first month, during which rates of post-discharge mortality may exceed initial inpatient mortality.2 3 The range of illnesses to which patients are susceptible is extremely broad.4 This period of heightened and generalized vulnerability to a broad spectrum of conditions has been called the post-hospital syndrome.5 Yet risk after discharge from hospital remains incompletely characterized. Previous research has largely focused on calculating rates of readmission and death within the first month after discharge (for example, 30 day readmission rates).2 6–8 In contrast, a few studies have compared relative differences in risk after hospital stay among different patient groups admitted with the same condition.9–11 Yet none of these studies have characterized patients’ absolute risks of readmission and death or the extent to which these risks change with time over the full year after hospital discharge. We therefore do not know when these risks are highest after
What is already known on this topic Patients are at high risk for readmission to hospital and death in the month after discharge However, little is known about how these risks dynamically change over time for the full year after hospitalization Accurate information on risk is needed for patients and hospitals to set realistic goals and plan for appropriate care
What this study adds Risk declines slowly after discharge from hospital and is increased for months compared with people who have not been hospitalized Specific risk trajectories vary by discharge diagnosis and outcome Patients should remain vigilant for deterioration in health for an extended time after discharge from hospital Health providers can use knowledge of absolute risks and their changes over time to better align interventions designed to reduce adverse outcomes after discharge from hospital with the highest risk periods for patients the bmj | BMJ 2015;350:h411 | doi: 10.1136/bmj.h411
1
RESEARCH discharge; when they become stable over time, with minimal day to day change; and the extent to which they are increased compared with a general elderly population. We also do not know if risk varies by admitting condition or outcome. This knowledge of absolute risks and their changes over time is critical to defining the period in which post-hospital syndrome persists and can help inform patients and their healthcare professionals about the timing of vulnerability after hospital discharge. This information is needed to set realistic expectations and goals for recovery after discharge. These data can also help hospitals to more efficiently align interventions designed to reduce adverse outcomes after hospital stay to the highest risk periods for patients. Several specific questions need to be addressed. Firstly, are the absolute risks of hospital readmission and death increased beyond 30 days after discharge, indicating that patients and providers should maintain high vigilance for deterioration in health beyond the initial month after hospitalization? Secondly, do the risks of readmission to hospital and death decline at different rates over time, suggesting that these outcomes result from different factors and therefore require different interventions to be most effectively reduced? Thirdly, does the period of increased risk differ by the initial condition triggering admission to hospital, implying a potential benefit of tailoring the duration and intensity of follow-up to the index diagnosis? Fourthly, do risks of readmission to hospital and death eventually approach plateau periods of relative stability, suggesting that patients may have entered a new phase of recovery with reduced vulnerability? Finally, is the risk of hospital admission and death after discharge noticeably higher than among the general elderly population for the full year after discharge from hospital, indicating that acute illness and hospital stay have longstanding affects on major adverse outcomes? Accordingly, we define absolute risks of first readmission to hospital and death and their changes in the year after discharge among a national cohort of Medicare fee for service beneficiaries surviving hospitalization for heart failure, acute myocardial infarction, or pneumonia. These three conditions are common reasons for hospital admission among older adults12 and have been the focus of public reporting.13 Knowledge of explicit risk trajectories can help patients and physicians set realistic goals and can help hospitals to more efficiently align the duration and intensity of follow-up care with each patient’s condition specific risk for readmission to hospital and death.
Methods Study sample We used Medicare standard analytic and denominator files to identify all admissions to acute care hospitals from 2008–10 with a principal discharge diagnosis of heart failure, acute myocardial infarction, or pneumonia. Cohorts were defined using international classification of diseases, ninth revision, clinical modification (ICD-9-CM) codes identical to those used in the publicly 2
reported readmission and mortality measures14–18 from the Centers for Medicare & Medicaid Services (see supplementary table 1). We included admissions to hospital among patients aged 65 years or older. We excluded patients with in-hospital death, less than one year post-discharge enrolment in Medicare fee for service in the absence of death, transfer to another acute care facility, discharge against medical advice, and uncertain vital status. As with the federal readmission measures,14–18 we used all index hospital hospitalizations across three years of study for analyses of readmissions to hospital. We restricted analyses of death to one random hospitalization per patient over the three year period to avoid repeated measurement of patients who died within one year of multiple admissions for heart failure, acute myocardial infarction, or pneumonia.
Study endpoints For the year after hospitalization, we identified the occurrence of first readmission to hospital and death on each day after discharge. As with the federal readmission measures,15 16 18 we only included readmissions to short term acute care hospitals and excluded all planned readmission to hospitals based on the presence of specific ICD-9-CM procedure and principal diagnosis codes.19 We did not consider transfers to other hospitals on the day of discharge or the next day after discharge to be readmission to hospitals.15 16 18 Comparator population To compare the risks of readmission and death after hospitalization for heart failure, acute myocardial infarction, or pneumonia with the risks of hospital admission and death in the general elderly population, we constructed a comparator population using the 2009 Medicare denominator and provider analysis and review files. The Medicare denominator file contains information on beneficiaries’ enrolment status in Medicare fee for service, date of birth, and date of death. The Medicare provider analysis and review file contains information on inpatient hospital admissions for enrolled Medicare fee for service beneficiaries, including the principal discharge diagnosis, date of admission, and date of discharge. Our comparator population included all Medicare fee for service beneficiaries aged 65 years or older on 1 January 2009 with at least 12 months of enrolment in fee for service Medicare in the absence of death. Outcomes Daily risks of readmission and death We estimated the daily risks of first readmission to hospital and death by day (1–365) after hospitalization for heart failure, acute myocardial infarction, or pneumonia. To illustrate risk trajectories, we identified the length of time required for daily risks of first readmission to hospital and death to each decline 50% from its maximum value after discharge. We also characterized the length of time required for daily risks of first readmission to hospital and death to approach plateau periods of minimal day to day change by calculating the doi: 10.1136/bmj.h411 | BMJ 2015;350:h411 | the bmj
RESEARCH number of days required for the daily change in risk of each to decline 95% from its maximum daily decline after discharge.
Relative risks of admission and death We characterized the extent to which the risks of hospital admission and death are higher among patients recently discharged from hospital compared with the general elderly population. This was done by calculating the cumulative incidence of readmission to hospital and death after hospitalization for heart failure, acute myocardial infarction, or pneumonia and comparing these results with the cumulative incidence of hospital admission and death among all beneficiaries in the Medicare fee for service comparator population. Statistical analyses Daily risks of readmission and death We fit separate survival models for the risk of first readmission to hospital and death after hospitalization for heart failure, acute myocardial infarction, or pneumonia. The analytic approach differed based on the presence or absence of competing risk. In models that estimated the daily risk of first readmission to hospital, we considered death before first readmission to hospital as a competing risk and therefore calculated the subdistribution hazard—an unconditional hazard—that was derived from the cumulative incidence function by Fine and Gray and corrects for competing risk.20 This approach allows the estimation of unconditional risk after consideration of competing risks. We censored data at planned readmission19 or at one year after the index hospitalization, whichever occurred first. In models that estimated the daily risk of death, there was no competing risk and we censored data at one year after the index hospitalization. In the absence of competing risk, we calculated hazard estimates for death using the life table method. We used Gray’s test to compare the cumulative incidence of first readmission to hospital and its corresponding hazard across cohorts with heart failure, acute myocardial infarction, and pneumonia.21 To compare the cumulative incidence of death and its corresponding hazard across the cohorts, we used the log rank test. We used the bootstrap method with 2000 iterations to construct 95% confidence intervals for the time required for the daily risks of first readmission to hospital and death to decline 50% from their maximum hazards after discharge for each of the three index conditions. To characterize the daily change in risk of first readmission to hospital and death with time after hospital discharge, we calculated differences in kernel-smoothed hazard estimates22 between each day and the preceding day. For each day after the maximum hazard, we divided the daily change in risk by its maximum daily decline after discharge. We used the bootstrap method with 2000 iterations to construct 95% confidence intervals for the number of days required for the daily change in risk to decline 95% from its maximum daily decline after discharge. the bmj | BMJ 2015;350:h411 | doi: 10.1136/bmj.h411
Relative risks of admission and death We calculated the one year cumulative incidence of hospital admission and one year cumulative incidence of death among all beneficiaries in the Medicare fee for service comparator population in 2009. We prorated these results by day and compared them with the cumulative incidence of hospital readmission and cumulative incidence of death by day (1–365) after discharge from index hospitalization for heart failure, acute myocardial infarction, or pneumonia. We did this by calculating the relative risk of hospital admission and death between study cohorts and the Medicare fee for service comparator population over the first 30, 60, 90, 180, and 365 days after discharge. To make study and comparator populations more similar, we directly standardized cumulative incidence by age, sex, and race. We used three age categories (65–74, 75–84, ≥ 85), two sex categories, and three race categories (white, black, other). We used the Fine and Gray method to derive the cumulative incidence function of first readmission to hospital for each age-sexrace stratum in the heart failure, acute myocardial infarction, and pneumonia cohorts.20 We derived the cumulative incidence of death for each age-sex-race stratum using the life table method. We directly standardized the study populations to the Medicare fee for service comparator population to calculate the age-sexrace standardized cumulative incidence of hospital admission and death. AFH, ZL, and HL conducted analyses using SAS 9.3 (SAS Institute, Cary, NC). Results For readmission analyses, we included 1 462 453 hospitalizations for heart failure (4735 hospitals), 561 926 for acute myocardial infarction (4423 hospitals), and 1 125 234 for pneumonia (4767 hospitals). Supplementary figure 1 shows the reasons for excluding hospitalizations. The cohorts were comprised, respectively, of 972 339, 516 380, and 951 084 unique patients who were used in mortality analyses. The comparator population of Medicare fee for service beneficiaries included 27 764 699 people. Table 1 lists the characteristics of the cohorts. Within one year of discharge, readmission to hospital and death, respectively, occurred following 67.4% and 35.8% of hospitalizations for heart failure, 49.9% and 25.1% for acute myocardial infarction, and 55.6% and 31.1% for pneumonia. The daily risk of first readmission to hospital 30 days after hospitalization for heart failure, acute myocardial infarction, or pneumonia was 0.7%, 0.4%, and 0.4%, respectively. The daily risk of death 30 days after hospitalization for heart failure, acute myocardial infarction, or pneumonia was 0.2%, 0.2%, and 0.2%, respectively. Table 2 presents the daily risks of first readmission to hospital and death at 7, 15, 30, 60, 90, 180, and 365 days after discharge. Although daily risks of first readmission to hospital and death both declined with time, the decline was relatively slower for first readmission to hospital for all three conditions (figure 1). After hospitalization for 3
RESEARCH
Table 1 | Personal characteristics of study cohorts and comparator population. Values are percentages unless stated otherwise Study cohorts Heart failure Characteristics
Mean (SD) age (years) Women Race: White Black Other
Acute myocardial infarction
Pneumonia
Readmission (n = 1 462 453)
Death (n = 972 339)
Readmission (n = 561 926)
Death (n = 516 380)
Readmission (n = 1 125 234)
Death (n = 951 084)
Medicare fee for service population (n = 27 764 699)
80.5 (8.2)
80.7 (8.2)
78.7 (8.4)
78.6 (8.4)
80.3 (8.2)
80.3 (8.2)
75.5 (7.9)
55.9 (n = 817 059)
56.2 (n = 545 923)
50.0 (n = 280 986)
50.1 (n = 258 883)
55.1 (n = 620 015)
55.7 (n = 529 504)
56.7 (n = 15 731 675)
83.0 (n = 1 213 963) 12.5 (n = 182 593) 4.5 (n = 65 897)
84.9 (n = 825 341) 10.8 (n = 105 173) 4.3 (n = 41 825)
88.0 (n = 494 226) 7.7 (n = 43 033) 4.4 (n = 24 667)
88.2 (n = 455 462) 7.5 (n = 38 612) 4.3 (n = 22 306)
88.8 (n = 998 925) 6.7 (n = 74 854) 4.6 (n = 51 455)
88.7 (n = 843 191) 6.8 (n = 64 694) 4.5 (n = 43 199)
86.7 (n = 24 074 306) 7.6 (n = 2 118 292) 5.7 (n = 1 572 101)
heart failure, acute myocardial infarction, or pneumonia, risk of first readmission to hospital was highest on day 3, day 2, and day 2 after discharge, respectively, and declined 50% after day 38, day 13, and day 25 (Table 3). Risk of death was highest on day 1 for patients with all three conditions and declined 50% after day 11, day 6, and day 10 after hospitalization for heart failure, acute myocardial infarction, or pneumonia, respectively. Daily risks of first readmission to hospital and death were different across the three index conditions (P or = 65 years. Epidemiol Infect 2008;136:232–40. © BMJ Publishing Group Ltd 2015
doi: 10.1136/bmj.h411 | BMJ 2015;350:h411 | the bmj
